Artificial intelligence is often hailed as making things faster, cheaper and easier, however in the employment law context it can actually drive-up legal fees once a dispute is underway.
In many respects, efficiencies can be found by using AI – for example, summarising documents, drafting correspondence, helping users feel more confident putting thoughts into words or proof-reading.
However, legal fees often increase as a result of how AI is being used by employees and employers – often before a lawyer ever becomes involved.
AI‑drafted emails: artificial ‘word salad’
One of the most common trends we’re seeing is clients using AI to draft emails both to us as legal advisers, and to the other party in a dispute.
On the surface, this seems helpful, because the correspondence initially appears polished, detailed, and structured. The issue is that – at best – AI does not know what matters legally and what doesn’t and – at worst – AI can give completely inaccurate information, such as legal claims that don’t exist or don’t apply in New Zealand.
A relatively simple piece of correspondence can arrive wrapped in several pages of background, commentary, emotional context, and tangential issues. From a lawyer’s perspective, that all needs to be read, digested and – often – clarified. Time is then spent working out what the actual issue is, rather than moving straight to providing quality legal advice.
In practice, this can mean more billable time being incurred just to triage the information. Ironically, a shorter email written “in your own words” is often far quicker – and cheaper – to respond to. As Mark Twain is often credited as saying – “I didn’t have time to write a short letter, so I wrote a long one instead.”
AI‑written personal grievances
AI is changing how employees raise personal grievances. Many employees are now using tools like ChatGPT to draft personal grievance letters that read more like legal pleadings than workplace complaints.
AI excels at expanding issues. When prompted with something like “write a personal grievance letter,” the output often includes unjustified disadvantage, unjustified dismissal – even if the employee is still employed – discrimination, bullying and breach of good faith – all of which can often detract from the validity of the claims.
The difficulty is that, once submitted, these personal grievances have to be taken seriously. Employers and lawyers can’t simply ignore parts of a personal grievance because they read as “AI‑generated” or exaggerated. Lawyers must assess each point, advise on risk, and often respond to claims that would never have been raised without AI’s prompting.
From a legal perspective, that creates immediate scope creep, or ‘word salad’, and therefore increased costs. Each allegation must be analysed separately. Advice becomes longer, correspondence becomes more detailed, and resolution becomes harder. What might have been a narrow issue suitable for early settlement can quickly turn into a complex dispute simply because of how it was framed.
Another subtle factor is that AI can make users feel legally fluent, even when they are not. That can entrench positions on both sides.
Employees may feel emboldened to pursue weak or peripheral claims. Employers may feel alarmed by the sophistication of a personal grievance and “over‑lawyer” the response. Both reactions can increase time and cost, even though the underlying facts have not changed.
A practical takeaway
When used strategically, AI can be a great tool and even reduce legal fees. The key is figuring out how to use it effectively. Leaving it to your lawyer to weed out the irrelevancies and repeated information comes at a cost.
If you are involved in an employment dispute, whether as an employee or an employer, consider using AI as a drafting assistant, not a legal strategist. For example, AI can be used to help prepare clear chronologies, summarise correspondence, or formulate focused questions for your lawyer.
In the world of AI over-saturation, our view is that focused issues, clear communication, and sound (human!) judgement remain the most effective ways to resolve problems efficiently and affordably, with a little help along the way from AI.
